Recommendation method and device based on deep learning
A technology of deep learning and recommendation method, applied in the field of recommendation, it can solve the problems that the recommended objects are not screened, the recommendation with high accuracy and high satisfaction cannot be achieved, and the recommendation efficiency is low, so as to achieve good noise immunity and effectiveness. Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Example Embodiment
[0055] Example 1
[0056] like figure 1 As shown, a recommended method based on deep learning, including the following steps:
[0057] S110, acquire multiple user portraits and multiple commodity properties, and lock the target user according to the plurality of user portraits;
[0058] S120, extract the display of the target user and the display characteristics of the plurality of commodities, and then generate a list of recommendations;
[0059] S130, using a depth learning model to learn the hidden features of the target user and the hidden features of the multiple commodity properties, and predict the target user's score of the recommended list according to the syndrome;
[0060] S140, the product of the recommendation list is pre-recommended to determine whether the target user receives it. If so, the score of the recommendation list is recommended in priority.
[0061] In Embodiment 1, a plurality of user portraits are acquired, and the plurality of user portraits include us...
Example Embodiment
[0062] Example 2
[0063] like figure 2 As shown, a recommended method based on deep learning, including:
[0064] S210, acquire multiple user portraits and multiple commodity properties, the user portrait includes behavioral features and preference features, the commodity attribute including product base attributes and product evaluation;
[0065] S220, set the multi-dimensional screening according to the behavior characteristics and preference features, and lock the target user according to the multi-dimensional screening;
[0066] S230, extract the display of the target user and the display characteristics of the plurality of commodity properties, and then generate a list of recommendations;
[0067] S240, lecture with a deep learning model to learn the hidden species of the target user and the hidden features of the multiple commodity properties, and predict the target user's score of the recommended list according to the synscies;
[0068] S250, the product of the recommendat...
Example Embodiment
[0070] Example 3
[0071] like image 3 As shown, a recommended method based on deep learning, including:
[0072] S310, acquire multiple user portraits and multiple commodity properties, and lock the target user according to the plurality of user portraits;
[0073] S320, the behavior characteristics of the target user are collected and the browsing records and search records of the plurality of goods are constructed, and the preference model of the target user is constructed;
[0074] S330, simultaneous collecting display features of the plurality of commodities attributes, and finds approximate goods according to the characteristics of the characteristics;
[0075] S340, a commodity cache stored in the approximate product and the preference model into a recommended list;
[0076] S350, using a deep learning model to learn the hidden species of the target user and the hidden features of the plurality of commodity properties, and predict the target user's score for the recommended...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap